Polychlorinated biphenyls (PCBs) contaminate seafood and freshwater fish worldwide. Current risk assessment is based on immunotoxicity in rodents, although they differ from humans in regard to immune system development; epidemiological confirmation is lacking. We have carried out a pilot study in the Faroe Islands and found that prenatal exposure to dioxin-related PCB congeners appears to affect the antibody response to childhood vaccinations, possibly mediated via a toxic effect on prenatal T-cell programming. This North Atlantic fishing community is unique and highly suitable for prospective population-based studies. Average PCB exposures vary more than 100-fold within the population, and confounding variables other than marine food contaminants are of limited concern. We now propose to examine a birth cohort of 547 children born over a 2-year period from 1 April 1998 at the Faroes. The children will first be examined just before the final vaccination booster against tetanus and diphtheria at age 5 years. An additional blood sample will be obtained 4-8 weeks later. The children will again be examined at approximately age 7 years. Outcome variables will include antibody responses to vaccines, acute-phase reactants, and questionnaire information on childhood infections. Stored maternal serum from late pregnancy, and serum from the children will be analyzed for PCBs and other PHAH contaminants, and we will use a novel method to assess the in vitro activation of the arylhydrocarbon receptor (AhR or 'dioxin' receptor) by the contaminants present in serum. Mercury will be measured in cord blood, postnatal blood and hair. Statistical analyses will include multiple regression analysis, supplemented by mixed model analyses, structural equation models, and determination of benchmark doses of individual contaminants, their mixtures, and integrated exposure measures. These results are expected to extend substantially the current documentation for PCB risk assessment. [unreadable] [unreadable]